Hi folks,
There appears to be a bug in boost::format, an unitialized value problem spotted when I was using valgrind on a program of mine.
The following test code illustrates the (admittedly minor?) problem:
Hi, sorry for the late reply, I dont check boost.user too often, and then it took for me to get it to accept my messages. I dont know valgrind, and I'm not sure I can figure out the nice logs you sent. it looks like a powerful tool, but does it say which variable is used uninitialised ? (and what compiler are you using format with, and does the sample code do anything weird there ?) I greatly doubt it is in format's code, I can find no possible reason for different behaviour between the 2 calls, within format's code. While in fact the stdlib's ostream code could very well do some tests and get in trouble with 0-valued int. Could it likely come from some code in the standard lib ? Also, I notice some lines in the valgrind report are left unmangled, could the analysis get mixed up at run-time for whatever reason, and detect an unitialised value by mistake ? regards, -- Samuel (boost::format author)